ahamilton454

interesting, so in the particular project where they not allowed to solve them, or what was the blocking mechanism that prevented a product hunt post?

This is quite an interesting approach. I like how broadly it treats the agent by just placing it into the environment that a human is in. Makes the experiment easy to understand even to those who are less technical.

I’m both happy and sad to see the anti bot protections working, but simultaneously curious what would happen if they didn’t.

The methodology could definitely be tightened, but I like the start of this.

Throwing a tiny little LLM at it helped me assess that it was far too slow for me to reasonably use at the scale I needed. So it didn’t really matter how accurate the prompt was. I was more just pointing out that I didn’t know if would be too slow without trying it. I maybe could have done some simple math in retrospect, but trying it out was easy enough

Yeah, Ive tried a bi-encoder, cross encoder and some small LLMs so far. I think I’ll do BERT soon too

Yeah exactly, I literally don’t know how to change my spec until I’ve gathered more data.

I was building a transaction classifier recently and I initially thought it would be a trivial “solved” problem. Throw transactions into a tiny local LLM, let it classify. But that approach was too slow, and not accurate enough. I didn’t know that though until I tried and then needed to change the spec.

I like that you point out that the cost to build software is still not 0. And in my expirence it’s further from 0 than I would expect. I often find myself thinking I can rebuild a project (or usually improve upon an existing one) in just a few days. And yet when it comes down to making anything well, it still takes time and iteration.

It’s a bit funny because I felt this way before coding agents as well, like you could clone something in just a few weeks. But in practice my expectations are rarely accurate.

Reading this on my eink Bigme Hibreak pro :).

I use the bigme hibreak pro as my primary device. Its an eink device but also just a full phone. Almost bought the boox palma, but held back because I only really want to carry a single device.

It doesn't always ask great questions, but even just the fact that it does makes me re-think what i am asking.

I definetly sometimes ask really specialized questions and in that case i just say "do the search" and ignore the questions, but a lot of times it helps me determine what i am really asking.

I suspect people with execellent communication abilities might find less utility from the questions

This is one of the reasons I really like deep research. It always asks questions first and forces me to refine and better define what I want to learn about.

A simple UX change makes the difference between education and dumbing users of your service.

I am somewhat in the same boat as this. At least i try to be super intentional about it. I replaced my iphone with the Hibreak bigme pro which is an e-ink phone.

Its not perfect. Ironically I still can doomscroll on it, but it is much bettter.

For youtube if you disable reccomendations, shorts eventually the feed disappears. I also use an extension called youtube block feed, which only allows me to see my subscriptions.

Its an imperfect firewall, but I am happy with it so far.

This was a pretty thoughtful piece and a pleasure to read. My gut reaction at first was that it was going to be some critque on children not reading due to generative ai (a cynial piece), but it was quite an interesting reflection on simply how reading is changing from the prepective of this author whose spent a his life deeply reading books

The ending had a nice flair of grandness to it as well.

The irony of an open source alternative to a fork of an open source project is hopefully not lost here

A stupidly simple CLI to store MCP configs. I need to change them all the time and move them from client to client which is annoying. Hopefully this solves my problems and maybe one of yours.

I’ve been struggling with a very similar feeling. I too am a manager now. Back in the day there was something very fulfilling about fully understanding and comprehending your solution. I find now with AI tools I don’t need to understand a lot. I find the job much less fulfilling.

The funny thing is I agree with other comments, it is just kind of like a really good stack overflow. It can’t automate the whole job, not even close, and yet I find the tasks that it cannot automate are so much more boring (the ones I end up doing).

I envy the people who say that AI tools free them up to focus on what they care about. I haven’t been able to achieve this building with ai, if anything it feels like my competence has decreased due to the tools. I’m fairly certain I know how to use the tools well, I just think that I don’t enjoy how the job has evolved.

I don’t think that is the core of this paper. If anything the paper shows that LLMs have no internal reasoning for math at all. The example they demonstrate is that it triggers the same tokens in randomly unrelated numbers. They kind of just “vibe” there way to a solution

A tool change can send a notification to the client, but the client chooses when to actually update the tools. This could take time and the LLM may not be aware of the new tool. I don’t think there is a concept of a “well-behaved” client since MCP is about giving flexibility of implementation.

I wouldn’t call this ergonomic. Alternatively, you could just notify the server when a user message is sent, and allow the server to adjust the tools and resources prior to execution of the agent (this is clearly different from the MCP spec).

On a separate note, what client are you using that supports notification, I haven’t seen one yet?

Server notifications are a bad way of implementing semantic retrieval on tools. There when would one update the tools? You can’t “trigger” an event which causes a tool change without some hacky workarounds

It’s actually just a Forbes marketing ploy. Me and my company made it on the list and they gave us an interview and the guy that interviewed us asked “what is ChatGPT?” And he was supposed to be head of like AI at Forbes. I actually think they somehow make money from it by hosting conferences and stuff. Either way I always was sus of it

I don’t understand the advantage of having the transport protocol be HTTP/SSE rather than studio especially in this case when it’s literally running locally.

Not much to say other than this is fun and awesome! Haven’t taken a detailed look at all, but in surface level, very enjoyable!

I got pretty hungry one day and decided that copilot should solve this problem. LMK if you have any questions about tech behind it!

Here is the vscode link to the extension if you want to try it yourself:

https://marketplace.visualstudio.com/items?itemName=buildwit...

I can agree this is a huge problem with large APIs, we are doing it with twilios api and it’s rough

Hey this looks pretty interesting. I saw that you guys are a YC company, how do you intend on making money deploying a protocol?

A really brief note what companies are getting into the LLM Extensibility space, and what each of their benefits and downsides are.

This is the A7 which is an older version. I would make sure it still has 4G support otherwise no carriers will support it in the US

I have done that and it is very nice. I also use this chrome extension https://chromewebstore.google.com/detail/unhook-remove-youtu...

It feels like an iphone 8 keyboard that is about to die. It isn't the quickest thing in the world, but good enough to respond to texts and emails.

about 3 days (maybe 4 days) of normal usage

about 3 days long (maybe 4)